Significant postoperative bleeding is a problem well known in any cardiac surgical center. It has to be expected in 5–10% of patients depending on the type of surgery and potential pre-existing risk factors [3]. It has been demonstrated that transfusion requirements have a considerable impact on postoperative outcome and are associated with increased mortality [17].
Optimally, treatment algorithms would be based on coagulation tests with rapid results, independent of heparin and capable of assessing the current function of the different coagulation system components. Theoretically, modified thrombelastometry systems designed as a “point-of-care” test seems well suited and have been proven to further reduce transfusion requirements in a small prospectively randomized trial compared to an algorithm solely based on standard tests [18].
An advanced “point-of-care” system (ROTEM®) has been introduced that allows for simultaneous testing in four chambers with different activation agents (“intrinsic” / “extrinsic”), platelets blockade and aprotinin or heparinase addition. The clinical introduction of this thrombelastometry-based blood component treatment has been demonstrated to reduce transfusion requirements and costs [11, 15, 19, 20]. The aim was to study a thrombelastometry-based algorithm in comparison to a protocol based on classic coagulation tests known to be superior to “empiric” treatment.
Overall we observed comparable transfusion requirements between the ROTEM®-guided and the classic coagulation test guided group regarding RBCs, FFPs, platelets and pooled factors (PPSB). Our ROTEM®-guided blood component treatment algorithm was at least as effective and safe as protocols based on classic coagulation tests. However, distribution blood products differed between both groups. ROTEM® patients received more fibrinogen but significantly less aprotinin, as already stated in a recent study [11].
There was a clear trend towards less 24 h-bleeding visible in the ROTEM® group, also more patients were on coumadin treatment preoperatively and heparin dosages were significantly higher in the ROTEM® group. When analyzing the subgroup of “high-risk” patients with long CPB-times known to be prone to diffuse coagulopathy ROTEM® guided treatment resulted in significantly less 24 h-drainage loss and an improved 5-year survival. It seems that the more specific approach of the ROTEM® based algorithm resulted in less bleeding. In addition, the ability to deliver the specific treatment faster than with standard tests may have contributed to the observed benefit. The ROTEM® protocol requires 10 min until the results allow for a decision if additional protamine, FFPs or platelets are required and after 30 min measurements will guide platelets, fibrinogen or antifibrinolytic agent therapy. In contrast, at least in our center, results of standard coagulation test are rarely available in less than one hour.
Rate of re-thoracotomy was not different between the two groups. To differentiate between surgical bleeding and diffuse coagulopathy is always a difficult task. Standard coagulation tests have been shown to be of no help at all in this scenario: negative predictive accuracy has been reported with 50% - basically the same as guessing [5]. In contrast, ROTEM® is more accurate and allows for a negative predictive value (excluding diffuse coagulopathy) of 82%, however positive prediction is less [5, 21]. In conclusion, ROTEM® might be helpful in the assessment of surgical versus diffuse bleeding but overall accuracy seems to be not sufficient. Thus, the decision for re-thoracotomy in our series was based on clinical judgment and only partially based on ROTEM® or standard coagulation test. As in only 21% of control and in 17% of ROTEM® patients a surgical bleeding could be identified, a substantial number of bleeding events could have been avoided with a better decision making protocol. Ideally, specific predictive values derived from ROTEM® or other coagulation tests might become available in the future facilitating a more evidence based approach if or if not to perform a re-thoracotomy [22, 23].
Avarage costs for the blood products used in this study were estimated according to Spalding et al. [11]. Consistent with previous studies we could confirm in our prospective randomized trial a cost-reduction when using a specific ROTEM® based treatment algorithm (Tables 6 and 7). Cost-savings were not counterbalanced by the additional costs for thrombelastometry testing (Fig. 2). Thus, it has to be considered that the implementation of such a ROTEM®-based protocol requires substantial human training and a dedicated and highly motivated ICU team as the work-load is increased due to the requirements to manually perform the ROTEM® tests on the ICU in comparison to simply sending a blood sample to the laboratory for traditional coagulation tests.
Another issue of concern is the accuracy and reproducibility of a “point-of-care” test performed manually by “non-specialized” physicians outside the laboratory environment. However, for the ROTEM® device sufficient reproducibility and stability of measurements have been reported [24,25,26]. The use of fully-automated systems like ROTEM® sigma need to be disseminated more widely to stimulate a broader use of this beneficial technology.
We observed a beneficial effect of the ROTEM® algorithm predominantly in the subgroup of “high-risk” patients with prolonged CPB-times. However, we have to consider that due to the study design ROTEM® was compared to a well-structured and evidence based standard transfusion protocol.
In this study, there was an overall trend towards improved survival at 30-days, however in the subgroup of “high-risk” patients known to be prone to diffuse coagulopathy a significant reduction in 5-year mortality was proven in the long-term follow-up. Standard treatment algorithms have been proven to significantly reduce transfusion requirements in comparison with “empiric” therapy [1, 5, 7, 27]. Hence, the observed benefit of a ROTEM®-guided protocol has been shown to be effective and safe in a prospective randomized trial in cardiac surgical patients that suffer postoperative bleeding and secondly a further reduction of bleeding, costs and mortality could be demonstrated. In our opinion, “empiric” blood component therapy should not be used in clinical practice. For regular patients both, standard and ROTEM®-based specific treatment algorithms have been shown repeatedly superior providing a true clinical benefit for the patients [1, 5].
With regard to a recent updated meta-analysis by Serraino and collegues [28] ROTEM-guided algorithms lead to a significant reduction in transfusion of RBCs, FFPs, platelets and the rate of severe acute kidney injury compared to CONTROL groups. There was also an improvement seen in mortality, number of reoperations for bleeding, ventilation times, shorter ICU length of stay and hospital stay, but none of them significantly reduced. The authors concluded that viscoelastic testing lacks clinical effectiveness with only weak evidence and low predictive accuracy for coagulopathic bleeding. However, a treatment algorithm based on “point-of-care” 4-chamber ROTEM® seemed to be at least as effective as standard therapy with improvement in a broad range of relevant clinical parameters. Furthermore, our incidence of postoperative acute kidney injury was rather low for a sample of patients with significant bleeding. Interestingly, our results are supported by newer studies, who demonstrated that a low nadir hematocrit (cutoff value of about 24%) was inversely associated with acute kidney injury [29,30,31].
A ROTEM® guided protocol seems to be capable to further reduce bleeding, costs and mortality as it allows for a highly specific and fast therapy tailored to the functional coagulation status of the individual patient and might be especially beneficial in “high-risk” patients prone to bleeding complications [14].
Limitations
The major limitation of this trial is that patients with aortic valve replacement as well as with combined CABG were included. Thus, our data might include a considerable preoperative inhomogeneity regarding the incidence of von-Willebrand-Syndrom. In addition, limited evidence is available in regard to the structure of the ROTEM® algorithm and baseline reference values [32] which might not allow comparability between “regular” and cardiac surgical patients after CPB. Furthermore, the small number of patients as well as surgical bleeding, that cannot be treated based on thromboelastometry, may have biased our results. However, it is still a unique randomized trial assessing ROTEM® effectiveness in cardiac surgical patients in case of postoperative bleeding. Furthermore, the rate of inadequate platelet response was not assessed in our study, but might be equally distributed within the groups due to randomization.
Within the duration of the trial Trasylol had been withdrawn from the marked. Subsequently, aprotinin was replaced by tranexamic acid. However, this change in treatment affected both groups equally and in addition tranexamic acid has been shown to be as effective as aprotinin in several trials [33].
